The Conundrum of Education

Duolingo, Robot Teachers, and VR at School. What Does The Future of Education Look Like?

This one of a series of posts written by FOV Ventures Entrepreneur-In-Residence Tiago Correia that take a deep dive into various topics we’ve been exploring at FOV Ventures, taking a look at their history and some of the new startups building within the area.

Education is a big space. For the purpose of this article, we are excluding anything before typical mandatory school in most jurisdictions, and anything done by employers or in the context of enterprise. This article looks at:

  • Why isn’t Education More Gamified?

  • The Startups Reinforcing Education’s Inherent Gaming Nature

  • VR and the Classroom

  • AI Education, and

  • The Joy of Learning.

There’s a lot to cover, but we hope you enjoy this topic deep dive, and look at the startups working in this field.

500 Million of us have had Duolingo installed on our phones at some point in time.

You might be one of the users who don’t use it regularly, but let me take a moment to step out of the 3rd person and tell you something: I am one of the 83 million monthly active users - an impressive 16.6% active user base.

Keeping us addic… engaged! Source: Duolingo

Sure, it pales in comparison to TikTok’s 31.91% - an app in which Duo, Duolingo’s owl mascot, is, coincidently, a star - but let’s not forget: we’re talking about a learning app. Learning.

Universally groaned as ‘not the funnest of activities’, the learning mobile app category has one of the lowest retention rates. Yet, 83 million of us are boasting to you about how long our streak is.

The piece almost writes itself: “Duolingo uses gamification to increase usage”. Points, badges, status, leagues, leaderboards, everything: they do it extremely well. And this coming from someone who slaved away an entire Sunday just to get the coveted #1 spot in the Diamond League, all that for a virtual badge. Mind you, slaving away at school for good grades has some semblance to virtual badges.

This begets the key question: Education, itself, is already inherently gamified; so why isn't it as ‘sticky’ as Duolingo?

We don’t need to go deep into the history of education… it’s been very homogenous. Sure, we’ve evolved from a one-to-one tutoring method to a one-to-many method, primarily based in the classroom. Despite global evidence favoring practical learning methods like Montessori, most educational systems worldwide still cling to traditional formats of lectures, book study, and high-stakes exams. This uniformity persists, curiously unchanged, in the face of clear benefits of hands-on learning. It’s easy to understand why so much is at stake, yet why nothing changes:

  • It all costs a lot of money: To equip classrooms, to retrain educators, to rewrite manuals. Government curriculums tend to be static because there’s a treasury incentive to keep costs low.

  • Governments would need to be on board: because if any improvements remain in the confines of private schools, for example, only 6% of the UK population would benefit - typically the percentage which already has an outsized comparative advantage.

  • Which poses a significant risk: There is a long and established correlation between education and economy. To put it bluntly, the higher the average years of schooling in a country, the higher its GDP. Anything to disrupt this balance is taken very suspiciously. 

And so, the status quo remains in an eternal catch 22. One in which 80% of employers believe graduates aren’t work-ready when entering employment; and one in which three in five young people believe they are ill-equipped for the job market. Add to that, that anyone indeed being equipped, has a 65% chance they’ll be employed for jobs that don’t exist yet, and you have the conundrum in which education exists today.

Much like the future of work, it took a global pandemic to force us to think beyond the typical confines of education. The damages reportedly incurred because of COVID to what UNESCO calls a “lost generation” of learners, are earth-shattering: the World Bank estimates them at 10% of the global GDP. At the same time, educators and learners are struggling to agree how AI tools, such as GPT, can actually play a part in formal education - but both know, the genie isn’t going back in the bottle.

Reinforcing Education’s Inherent Gaming Nature

The times when pointing and clicking was actually gaming. Source: The Magic School Bus Explores The Human Body

One of the OGs of the space has to be Norway’s quiz creator tool, Kahoot. It is so simple that it’s almost impossible to see any fault of it being seamlessly used in the classroom. And so it has.

It’s this same simplicity that has allowed Brainscape’s flashcards to become a companion to many students wanting to find a more engaging way to memorise studies. It is tangentially part of a slew of ‘gaming’ apps, from Lumosity to Elevate, which use flat colours and a sleek design on top of basic visual reward mechanics to enhance your cognitive skills - the so-called brain game apps - the equivalent of which, targeted at young learners, like IQClub, aim to be development companions. 

“But mom, I’m not playing, I’m LEARNING!” Source: Roblox / Museum of Science in Boston

But gaming+education is not just about creating games that teach. This is where the brilliant Kinjo has flipped the model. We’ve all learned through the games we play - so wouldn't it be great if parents could know that their kids were actually being helped by games?

Kinjo’s whole MO is also about finding cognitive development or even educational material in existing games - Roblox - and rewarding them when they play approved games that make the most out of their screen time. As you start moving into the realm of games, you also bring elements which are not defined by the bounds of a 2D page from a textbook, which is the same as…

Extending the Reality of The Classroom

What we all wish VR learning was like. Source: T. A. Furness / SAGE Publications.

The use of head mounted displays (/VR) has been trialled, tested and implemented across school systems. There is overwhelming data proving that when good learning content is developed it can be more engaging, particularly for learners who don’t fare well with the classroom model.

The trick here becomes doing so with content that can fit the curriculum, again, to make this not just the confines of private/alternative education. This is where companies like PrismsVR are excelling, using VR to teach maths and its concepts, in ways that we wished we had when we were students.

Naturally, maths is a hard challenge to tackle. But if a solution really wants to be integrated, to make use of the equipment, it must do so across more than one field. Exactly why Estonia-based Futuclass bundles the hardware with educational content in a wide range of topics.

Collectively learning through the power of imagination… and AR. Source: Brainspark

But the reality is that VR headsets are expensive, and individual. Which means that any sunk cost can only be divided by a max utilisation rate of one person at a time. This is where the current state of AR has an advantage. Currently the confines of tablets - predominantly cheaper and more commonly available hardware than headsets - AR allows for the same piece of content to be shared by two or more learners - increasing the utilisation rate and lowering costs.

While not being the only thing that British startup Brainspark Games is doing, this is where they have an interesting advantage when speaking to school officials and parents. On a slightly different approach, Play Osmo, attach a camera dongle onto a student’s vertically raised tablet, so it checks answers and progress on written material, making sure schools have no reason to reject these new tools. As we venture onto 3D methods, particularly those which interact with the world around us, we open up the pandora’s box of…

AI and the role of machine-based educators

That’s it we’re doomed… or not. Source: Youtube

Textbooks are outdated in an era of dynamic information. However, their enduring presence in education is due to their resilience against revisionism and their ability to uphold the quality and rigor required by school systems, despite being impersonal and limited in content updates and storage capacity.

But let's face it: their role will diminish as AI models start to flourish in the classroom.

Your textbook doesn’t have a picture of ancient egypt? Get an AI model like Midjourney to create it. Not sure your notes from stats class really help you? Don’t fret, throw the problem into ChatGPT and ask it to explain to you how to solve it. One can see how quickly, not just the role of a textbook might fall to the wayside but how educators or even schools might fear some serious scope creep.

Please welcome your substitute teacher, Mr. Tech. Source: Century Tech

In fact, if you look at the cadre of assistant apps that have risen in the last few years, it’s possible. From reading practice assistants Ello and Amira, to language-learning chatbots Langotalk, Memrise, LangAI, to curriculum companion tools Eduaide, AI4edu, and Century. Even the the established players like Khan Academy and Udemy are getting in on AI-powered personalization.

Lest not forget, it was exactly with personalization efforts Duolingo’s green owl mascot, Duo, saying “Hi, it’s Duo” and directly encouraging users, that Duolingo increased DAUs by 5% and D14 retention by 7.2%. So why wouldn’t schools and educators use tools like TeachMate or OpenLearning to design bespoke teaching plans that play to their and their students’ strengths?

Even the fragmented state-by-state nature of education, hasn’t stopped startups from developing AI personalization tools - AIaaS tool Riiid, touts a plug and play approach to anyone anywhere building an educational platform. The reality is that if you open up the tools that have long been in the hands of curricula makers and provide that to the students, one can maybe even unlock…

The Joy of Learning: Boards, Kits, Bots, and All Manner of Toys

No, that’s not the MVP of T-800, that’s the lovable Cog. Source: MIT

As early as 1978, studies have proven the diversity in learning styles, with as much as 45% of the population being kinesthetic learners - those who prefer a tactile, physical mode of learning. This is obviously something that traditional school approaches, steeped in auditory and visual learning, have not encouraged. But add to the fact that, regardless of what kind of learning style you have as a base, based on a survey of college students, 67% of them reported they process information best by being physically active.

Now I’ve known VCs to bet on risky consumer preference shifts, yes. But I’ve yet to know any betting on human aptitude shifts. And as we see so many focus on lofty (read: big), noble (read: important) topics such as climate, well this is just the price to pay to play in education.

“You’re a coder, Harry”. Source: Kano

We all know this, because we remember old skool chemistry sets, plasticine or even the sadly defunct LEGO Mindstorms (RIP) - the kick ass programmable robots that you could control with a microcontroller. We remember the joy we all had through this discovery-based learning, sometimes even impacting career paths. The success of single-board computer projects like Raspberry Pi, has seen it be adopted to power some very weird projects indeed, but also educational companies like Kano (sadly also defunct).

But today feels like the dawn of a new cycle of startups, following in the footsteps of its programmable forebearers. Particularly in robotics, we’re seeing companies like Albebaran come out with the NAO, Wonder Workshop with the Dash, and Mobsya with the Thymio. And you can understand their appeal. Projects like the Cubetto, using the Logo turtle approach with movements on the floor, specifically brag up their ‘screenless approach’ to parents. But it’s not just programming that the world is fixated on, companies like BrainSTEM are indeed focusing on the good ol’curriculum, but just choosing a tactile way to do it.

As circuit boards, chips and other components keep getting cheaper and parents’ concerns on screen time keep getting higher, we are sure to see a definitive space to play for new startups wanting to redefine ‘play to learn’.